Fix UUID Error 0XC002000B: Object Not Found on Windows Server

UUID 0xC002000B shows up when Windows tries to resolve an object reference and comes back empty. That's the whole error. The system asked for something by its unique identifier, the lookup succeeded mechanically, and the result was nothing. No object. No fallback. Just a failed resolution.

I've seen this on Server 2016 file servers after a storage migration, on 2019 domain controllers after an authoritative restore, and on 2022 nodes where someone deleted a DFS namespace target without cleaning up the referral. The trigger varies. The mechanism doesn't.

Cause 1: The object was deleted but a reference still points to it

What's actually happening here is a dangling identifier. Some component — a shortcut, a scheduled task, a DFS link, a certificate template, a Group Policy preference — stores the UUID of an object it depends on. When the object gets deleted, the reference doesn't vanish. It sits there, pointing at a UUID that no longer resolves. The next time anything touches that reference, you get 0xC002000B.

The most common real-world trigger: someone removes an AD user or computer account while a scheduled task, service logon, or SMB share ACL still references that account's SID or UUID. The account is gone. The reference isn't.

Find the broken reference. On a domain controller, run:

Get-ADObject -Filter {ObjectGUID -eq "00000000-0000-0000-0000-000000000000"} -IncludeDeletedObjects -Properties *

That's a placeholder GUID — swap in the UUID from your error log. If the object is in the Deleted Objects container, you know it was removed recently. Restore it with:

Restore-ADObject -Identity "CN=Broken,CN=Deleted Objects,DC=domain,DC=local"

If you can't restore it (tombstone lifetime expired, or it was never an AD object), the real fix is to hunt down the stale reference and delete or repoint it. Check Group Policy Preferences, scheduled tasks under C:\Windows\System32\Tasks, DFS namespace targets with dfsutil, and any shortcut with a stored SID.

Cause 2: Path or share reference is broken or unresolvable

UUID 0xC002000B also fires when a path-based lookup can't resolve. The identifier is valid, the object should exist, but the resolver can't reach it. This is the second most common scenario and it's almost always network or permission related.

Classic setup: a file share on Server 2019 has a DFS referral to a target on a different subnet. The target server is up, but the DFS root can't authenticate to it because the machine account password got out of sync. The UUID of the target is still registered in the namespace, so Windows tries to resolve it, fails at the authentication step, and reports the object as not found.

Check the referral chain:

dfsutil /pktinfo /display

You'll see the target listed with a status. If it says REFERRAL_UNAVAILABLE or similar, the UUID lookup is failing downstream. The fix is usually one of three things:

1. Reset the machine account password on the target: Reset-ComputerMachinePassword -Server DC01
2. Re-add the target to the namespace: dfsutil target add \\server\share
3. Fix the underlying SMB permission so the DFS root can read the target

On non-DFS shares, the same error appears when a symbolic link or mounted folder points to a volume that's offline. Windows stores the volume's UUID. The volume goes missing. Any access attempt through the mount point returns 0xC002000B. Check with mountvol and look for orphaned mount points.

Cause 3: Directory Services restore left orphaned UUID references

This one's nastier. After an authoritative restore or a non-authoritative restore on a domain controller, some objects come back with new UUIDs. Anything holding the old UUID — replication metadata, lingering objects, tombstoned references — becomes a dangling pointer. Server 2016 and 2019 DCs are especially prone to this because they don't aggressively clean lingering objects by default.

The real fix is to enable strict replication consistency and let the DCs clean house:

repadmin /regkey * +strict
repadmin /removelingeringobjects DC01 DC=domain,DC=local /advisory_mode

Run the advisory mode first. It tells you what would get removed without touching anything. Once you're confident, drop the /advisory_mode flag and let it run. Then force replication:

repadmin /syncall /AdeP

If the error persists after that, the reference isn't in AD at all. It's in a local component — typically the Certificate Services database or a Windows Internal Database on the DC. Check the certsrv logs for the UUID. If it's a certificate template reference, you'll need to reissue or delete the template.

Don't skip the advisory mode. I've watched someone run /removelingeringobjects without it and wipe out legitimate objects that were mid-replication. Wait for full convergence before you delete anything.

What usually isn't the cause

DNS. DNS gets blamed for everything, but 0xC002000B is not a name resolution failure. If DNS were the problem, you'd see 0x80070035 or 0x8007054B. UUID lookup failures are object resolution failures, which means the name resolved fine and the identifier didn't match anything on the other end. Don't waste time flushing DNS cache.

Also: don't restart the server first. The error survives reboots. It's a data problem, not a state problem. Rebooting just delays your troubleshooting by three minutes.

Quick reference

The pattern across all three: something is holding a UUID that no longer resolves. Find the holder, fix the reference. Don't chase the error — chase what's pointing at nothing.